CNN Visualization Sample¶
This samepl is from https://github.com/keras-team/keras/blob/master/examples/conv_filter_visualization.py BUT it is not work when I use udacity's anaconda environment. Because version of keras of anaconda is 2.1.6 but this sample have to use 2.2.0. And another error when I try to transfer layer block2_conv1 since the size of it is 128. Since I do some change in the code and hope it is helpful.
In [1]:
from __future__ import print_function
import numpy as np
import time
from keras.preprocessing.image import save_img
from keras.applications import vgg16
from keras import backend as K
# dimensions of the generated pictures for each filter.
img_width = 128
img_height = 128
In [2]:
def deprocess_image(x):
# normalize tensor: center on 0., ensure std is 0.1
x -= x.mean()
x /= (x.std() + K.epsilon())
x *= 0.1
# clip to [0, 1]
x += 0.5
x = np.clip(x, 0, 1)
# convert to RGB array
x *= 255
if K.image_data_format() == 'channels_first':
x = x.transpose((1, 2, 0))
x = np.clip(x, 0, 255).astype('uint8')
return x
def normalize(x):
# utility function to normalize a tensor by its L2 norm
return x / (K.sqrt(K.mean(K.square(x))) + K.epsilon())
In [3]:
model = vgg16.VGG16(weights='imagenet', include_top=False)
print('Model loaded.')
model.summary()
# this is the placeholder for the input images
input_img = model.input
# get the symbolic outputs of each "key" layer (we gave them unique names).
layer_dict = dict([(layer.name, layer) for layer in model.layers[1:]])
In [4]:
def output_img(layer_name, layer_dict):
kept_filters = []
layer_output = layer_dict[layer_name].output
min_len = min(layer_output.shape[-1], 200)
for filter_index in range(min_len):
# we only scan through the first 200 filters,
# but there are actually 512 of them
# print('Processing filter %d' % filter_index)
start_time = time.time()
# we build a loss function that maximizes the activation
# of the nth filter of the layer considered
layer_output = layer_dict[layer_name].output
if K.image_data_format() == 'channels_first':
loss = K.mean(layer_output[:, filter_index, :, :])
else:
loss = K.mean(layer_output[:, :, :, filter_index])
# we compute the gradient of the input picture wrt this loss
grads = K.gradients(loss, input_img)[0]
# normalization trick: we normalize the gradient
grads = normalize(grads)
# this function returns the loss and grads given the input picture
iterate = K.function([input_img], [loss, grads])
# step size for gradient ascent
step = 1.
# we start from a gray image with some random noise
if K.image_data_format() == 'channels_first':
input_img_data = np.random.random((1, 3, img_width, img_height))
else:
input_img_data = np.random.random((1, img_width, img_height, 3))
input_img_data = (input_img_data - 0.5) * 20 + 128
# we run gradient ascent for 20 steps
for i in range(20):
loss_value, grads_value = iterate([input_img_data])
input_img_data += grads_value * step
# print('Current loss value:', loss_value)
if loss_value <= 0.:
# some filters get stuck to 0, we can skip them
break
# decode the resulting input image
if loss_value > 0:
img = deprocess_image(input_img_data[0])
kept_filters.append((img, loss_value))
end_time = time.time()
# print('Filter %d processed in %ds' % (filter_index, end_time - start_time))
# we will stich the best 64 filters on a 8 x 8 grid.
n = 8
# the filters that have the highest loss are assumed to be better-looking.
# we will only keep the top 64 filters.
kept_filters.sort(key=lambda x: x[1], reverse=True)
kept_filters = kept_filters[:n * n]
kept_filters_maxlen = len(kept_filters)
# print(kept_filters_maxlen)
# build a black picture with enough space for
# our 8 x 8 filters of size 128 x 128, with a 5px margin in between
margin = 5
width = n * img_width + (n - 1) * margin
height = n * img_height + (n - 1) * margin
stitched_filters = np.zeros((width, height, 3))
# fill the picture with our saved filters
for i in range(n):
for j in range(n):
# img, loss = kept_filters[i * n + j]
# print(i * n + j)
if (i * n + j) >= kept_filters_maxlen:
break
img, loss = kept_filters[i * n + j]
stitched_filters[(img_width + margin) * i: (img_width + margin) * i + img_width,
(img_height + margin) * j: (img_height + margin) * j + img_height, :] = img
# save the result to disk
filename = layer_name + 'stitched_filters_%dx%d.png' % (n, n)
save_img(filename, stitched_filters)
return filename
In [6]:
# this section have to run couples hours in my mac.
# Don;t run it if you just want to see the result of vgg16
output_layers = ['block1_conv1', 'block2_conv1', 'block3_conv1', 'block4_conv1', 'block5_conv1']
img_files = []
for i in range(len(output_layers)):
f = output_img(output_layers[i], layer_dict)
print(f)
img_files.append(f)
In [9]:
from IPython.display import Image
Image(filename='block1_conv1stitched_filters_8x8.png')
Out[9]:
In [10]:
Image(filename='block2_conv1stitched_filters_8x8.png')
Out[10]:
In [11]:
Image(filename='block3_conv1stitched_filters_8x8.png')
Out[11]:
In [12]:
Image(filename='block4_conv1stitched_filters_8x8.png')
Out[12]:
